The section continues to investigate pharmacological and behavioral treatments of substance abuse and to explore combinations of treatments. We have already demonstrated the effectiveness of behavioral interventions (reinforcement for cocaine-negative urine samples) in large inner-city samples of intravenous polydrug abusers, and we continue to evaluate the best ways to apply the treatment. We have completed the follow-up phase of a study to address how a hypothetical methadone-maintenance clinic could best allocate its resources (both pharmacological and nonpharmacological) to reduce both heroin and cocaine abuse, if the clinic were to institute voucher-based contingency management. A novel contingency was used that reinforced abstinence from either drug while doubling reinforcer values for simultaneous abstinence from both. The total value of available vouchers was no greater than the amount used in previous trials targeting a single drug; instead, the amount was ?split? between cocaine and opiates. Vouchers were exchangeable for goods and services. The study used a 2x3 design in which 252 methadone-maintained outpatients were randomly assigned to a methadone dose condition (70 or 100 mg/day, double blind) and a voucher condition (contingent on cocaine-negative urines; ?split?; or noncontingent [vouchers given independent of urine test results]). The study lasted 27 weeks (Baseline assessment, 5 weeks; Intervention, 12 weeks; Maintenance, 10 weeks) with follow-up for 1 year. Urine results suggest that the dose increase reduced heroin use, but not cocaine use; though this result was expected, the present study provided one of the more unambiguous demonstrations to date. The effect of the split contingency on simultaneous abstinence from heroin and cocaine was modest: the Split 100mg group achieved a longer duration of simultaneous negatives than the control (NC 70mg) group, though many participants never provided even one simultaneous negative. For a split contingency to promote abstinence from cocaine and heroin simultaneously, a relatively high dose of methadone appears necessary but not entirely sufficient; an increase in overall voucher amount may also be required. Analyses of the follow-up data are ongoing. In a follow-up study, we are currently evaluating high-dose methadone combined with contingency management to maximize simultaneous abstinence from heroin and cocaine. We have included additional measures of medical status to examine potential health benefits of methadone maintenance. In the interest of cost containment and technology transfer, we conducted two pilot studies in which the reinforcers are lottery draws rather than vouchers, since work by others has shown that a lottery-based reinforcement procedure can be funded by donations from community manufacturers and merchants. The initial pilot study indicated that the lottery procedure generated substantial enthusiasm among study participants and is technically easier than the voucher program. In a further ongoing pilot study, we are comparing two different lottery procedures and two different densities of reinforcement. We may incorporate the lottery procedure into a more ambitious study in which a variety of low- or no-cost incentives are targeted toward a broad range of behavioral changes beyond reductions in drug use. Another current focus of our research derives from laboratory-animal data showing that stress-induced reinstatement of cocaine-seeking and/or heroin-seeking can be prevented with the alpha-adrenergic agonist clonidine, while cue-induced reinstatement of such drug-seeking can be prevented with the CB1 antagonist rimonabant. We have completed a dose-ranging study to determine how lofexidine can most safely be co-administered with methadone in humans. We are currently conducting a study in which methadone-maintained outpatients carry handheld computers throughout the day to provide real-time data on cravings for heroin and cocaine, lapses to drug use, and base rates of putative lapse precipitants. Once we have demonstrated the feasibility of this form of data collection in our population, we will be able to conduct clinical trials in which voucher-initiated abstinence is followed by maintenance on a medication specifically intended to prevent stress-induced or cue-induced relapse.